检索结果-内蒙古大学图书馆

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Jindong Xie Jiachen Liu Yupei Lin Jinbao Wang Xianxu Hou Linlin Shen Computer Vision Institute School of Computer Science and Software Engineering Shenzhen University National Engineering Laboratory for Big Data System Computing Technology Shenzhen University School of Information Engineering Guangdong University of Technology Guangzhou China Guangdong Provincial Key Laboratory of Intelligent Information Processing School of AI and Advanced Computing Xi’an Jiaotong-Liverpool University

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

The 3D generative adversarial network (GAN) inversion converts an image into 3D representation to attain high-fidelity reconstruction and facilitate realistic image manipulation within the 3D latent space. However, previous approaches face challenges regarding the trade-off between the reconstruction ability and editability. That is, reversing a real-world image to a low-dimensional latent code would inevitably lead to information loss, and achieving a near-perfect reconstruction using high-rate triplane representation often limits the ability to manipulate the image freely in the latent space. To address these issues, we propose a novel latent conditioning encoder-based framework with the alignment between the low-dimensional latent and high-dimensional triplane. A non-semantic guided editing strategy bridges the intrinsic relation between the latent condition and triplane generation, making it possible to edit the high-dimensional representation by latent manipulation. As a result, our method can achieve high-fidelity reconstruction and editing simultaneously by directly controlling the latent code. Experimental results demonstrate that our approach excels in reconstruction and editing quality compared to previous 3D inversion methods. Furthermore, our method can also edit even real faces with large poses and out-of-domain cases.

关键词： Bridges Three-dimensional displays Codes Signal processing Generative adversarial networks Vectors Acoustics Speech processing Image reconstruction Faces

来源：评论

学校读者我要写书评

暂无评论

DINOv2-powered Few-Shot Semantic Segmentation: A Unified Framework via Cross-Model Distillation and 4D Correlation Mining

arXiv

引用

arXiv 2025年

作者： Zhuo, Wei Tang, Zhiyue Xue, Wufeng Ding, Hao Shen, Linlin School of Artificial Intelligence Shenzhen University Shenzhen518060 China National Engineering Laboratory of Big Data System Computing Technology Shenzhen University China School of Biomedical Engineering Shenzhen University Medical School Shenzhen University China School of Computer Science and Software Engineering Shenzhen University China

Few-shot semantic segmentation has gained increasing interest due to its generalization capability, i.e., segmenting pixels of novel classes requiring only a few annotated images. Prior work has focused on meta-learning for support-query matching, with extensive development in both prototype-based and aggregation-based methods. To address data scarcity, recent approaches have turned to foundation models to enhance representation transferability for novel class segmentation. Among them, a hybrid dual-modal framework including both DINOv2 and SAM has garnered attention due to their complementary capabilities. We wonder "can we build a unified model with knowledge from both foundation models?" To this end, we propose FS-DINO, with only DINOv2’s encoder and a lightweight segmenter. The segmenter features a bottleneck adapter, a meta-visual prompt generator based on dense similarities and semantic embeddings, and a decoder. Through coarse-to-fine cross-model distillation, we effectively integrate SAM’s knowledge into our lightweight segmenter, which can be further enhanced by 4D correlation mining on support-query pairs. Extensive experiments on COCO-20i, PASCAL-5i, and FSS-1000 demonstrate the effectiveness and superiority of our method. Copyright © 2025, The Authors. All rights reserved.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Aspect category and sentiment polarity detection using a permutation language-based transformer model

引用

Procedia computer Science 2025年 258卷 4179-4189页

作者： Ashok Kumar Jayaraman Tina Ester Trueman Gayathri Ananthakrishnan Abirami Murugappan Erik Cambria Manikandan G Satanik Mitra HCLTech Engineering and R&D Services Chennai Tamil Nadu India Department of Computer Science University of the People California USA School of Computer Science Engineering and Information Systems Vellore Institute of Technology Vellore India Department of Information Science and Technology Anna University Chennai India College of Computing and Data Science Nanyang Technological University Nanyang Singapore School of Computer Science and Engineering Vellore Institute of Technology Vellore India Department of Management BITS Pilani India

Aspect-based sentiment analysis (ABSA) performs fine-grained analysis on text to determine a specific aspect category and a sentiment polarity. Recently, machine learning models have played a key role in ABSA tasks. In particular, transformer-based pre-trained models have achieved promising results in natural language processing tasks. Therefore, we propose a permutation based XLNet fine-tuning model for aspect category detection and sentiment polarity detection. Our model learns bidirectional contexts via positional encoding and factorization order. We evaluate the proposed permutation language model on three ABSA datasets, namely, SentiHood, SemEval 2015, and SemEval 2016. Specifically, we studied the ABSA tasks in a constrained system with a multi-class environment. Our result indicates that the proposed permutation language model achieves a better result.

关键词： Aspect-based sentiment analysis permutation language model deep learning transformers multiclass classification

来源：评论

学校读者我要写书评

暂无评论

Towards Lightweight Time Series Forecasting: a Patch-wise Transformer with Weak data Enriching

arXiv

引用

arXiv 2025年

作者： Wang, Meng Yang, Jintao Yang, Bin Li, Hui Gong, Tongxin Yang, Bo Cui, Jiangtao School of Computer Science Xi’an Polytechnic University China School of Data Science and Engineering East China Normal University China School of Computer Science and Technology Xidian University China Shanghai Yunxi Technology China The Shaanxi Key Laboratory of Clothing Intelligence China

Patch-wise Transformer based time series forecasting achieves superior accuracy. However, this superiority relies heavily on intricate model design with massive parameters, rendering both training and inference expensive, thus preventing their deployments on edge devices with limited resources and low latency requirements. In addition, existing methods often work in an autoregressive manner, which take into account only historical values, but ignore valuable, easy-to-obtain context information, such as weather forecasts, date and time of day. To contend with the two limitations, we propose LiPFormer, a novel Lightweight Patch-wise Transformer with weak data enriching. First, to simplify the Transformer backbone, LiPFormer employs a novel lightweight cross-patch attention and a linear transformation-based attention to eliminate Layer Normalization and Feed Forward Network, two heavy components in existing Transformers. Second, we propose a lightweight, weak data enriching module to provide additional, valuable weak supervision to the training. It enhances forecasting accuracy without significantly increasing model complexity as it does not involve expensive, human-labeling but using easily accessible context information. This facilitates the weak data enriching to plug-and-play on existing models. Extensive experiments on nine benchmark time series datasets demonstrate that LiPFormer outperforms state-of-the-art methods in accuracy, while significantly reducing parameter scale, training duration, and GPU memory usage. Deployment on an edge device reveals that LiPFormer only takes only 1/3 inference time compared to classic Transformers. In addition, we demonstrate that the weak data enriching can integrate seamlessly into various Transformer based models to enhance their accuracy, suggesting its generality. © 2025, CC BY-NC-ND.

关键词： Time series

来源：评论

学校读者我要写书评

暂无评论

Sentiment Analysis of Movie Reviews: Optimizing Predictions Through TF-IDF Vectorization

Sentiment Analysis of Movie Reviews: Optimizing Predictions ...

引用

Intelligent data Communication Technologies and Internet of Things (IDCIoT), International Conference on

作者： Mohd Irfan Anne Sreyas Venkatesam Hariharan Shanmugasundaram Mahendra Kumar Gourisaria Subrata Chowdhury Sudhansu Shekhar Patra Department of Artificial Intelligence and Data Science Vardhaman College of Engineering Hyderabad Telangana India Department of Artificial Intelligence and Data Science Vardhaman College of Engineering Hyderabad India School of Computer Engineering Kalinga Institute of Industrial Technology (KIIT) Deemed to be University Bhubaneswar India Dept. of Computer Science and Engineering Sreenivasa Institute of Technology and Management Studies(A) Chittoor Andhra Pradesh India

ISBN: (数字)9798331527549

ISBN: (纸本)9798331527556

Sentiment analysis model is aimed to accurately interpret movie reviews, addressing challenges like sarcasm and noisy data. By integrating multiple machine learning classifiers Naive Bayes, Random Forest, K-Nearest Neighbors, Logistic Regression, Decision Tree, and XGBoost—the model enhances sentiment prediction accuracy. The Term Frequency-Inverse Document Frequency (TF-IDF) method is used to extract features from a dataset of 50,000 movie reviews. Performance is evaluated using accuracy, precision, recall, and F1-score, with Logistic Regression achieving the best results, including an highest accuracy of 84.33 % and an F1-score of 0.847. In comparison to traditional methods, this multi-classifier approach offers a deeper understanding of movie reviews and its influence on audience decisions. The study demonstrates the potential of using these machine learning techniques to provide real-time, scalable sentiment analysis, which is crucial for filmmakers and marketers seeking insights into audience satisfaction and box office performance.

关键词： Sentiment analysis Logistic regression Accuracy Reviews Predictive models Motion pictures Feature extraction Real-time systems Noise measurement Random forests

来源：评论

学校读者我要写书评

暂无评论

RIDE: Enhancing Large Language Model Alignment through Restyled In-Context Learning Demonstration Exemplars

arXiv

引用

arXiv 2025年

作者： Hua, Yuncheng Qu, Lizhen Li, Zhuang Xue, Hao Salim, Flora D. Haffari, Gholamreza Department of Data Science & AI Monash University Australia School of Computer Science Engineering University of New South Wales Australia School of Computing Technologies Royal Melbourne Institute of Technology Australia

Alignment tuning is crucial for ensuring large language models (LLMs) behave ethically and helpfully. Current alignment approaches require high-quality annotations and significant training resources. This paper proposes a low-cost, tuning-free method using in-context learning (ICL) to enhance LLM alignment. Through an analysis of high-quality ICL demos, we identified style as a key factor influencing LLM alignment capabilities and explicitly restyled ICL exemplars based on this stylistic framework. Additionally, we combined the restyled demos to achieve a balance between the two conflicting aspects of LLM alignment—factuality and safety. We packaged the restyled examples as prompts to trigger few-shot learning, improving LLM alignment. Compared to the best baseline approach, with an average score of 5.00 as the maximum, our method achieves a maximum 0.10 increase on the Alpaca-eval task (from 4.50 → 4.60), a 0.22 enhancement on the just-eval-instruct benchmark (from 4.34 → 4.56), and a maximum improvement of 0.32 (from 3.53 → 3.85) on the MT-Bench dataset. We release the code and data at https://***/AnonymousCode-computerScience/RIDE. © 2025, CC BY.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Smart Web Application for Recommending the Selection of Equipment for Organizing Activities in Educational Institutions

Smart Web Application for Recommending the Selection of Equi...

引用

International ECTI Northern Section Conference on Electrical, Electronics, computer and Telecommunications engineering (ECTI-NCON)

作者： Tanarin Khongtuan Sakorn Mekruksavanich Master in Technology and Digital Data Managementc School of Information and Communication Technology University of Phayao Phayao Thailand Department of Computer Engineering School of Information and Communication Technology University of Phayao Phayao Thailand

ISBN: (数字)9798331543273

ISBN: (纸本)9798331543280

Higher education institutions strive to improve the quality of instruction and research to prepare students with essential 21st-century competencies. They promote academic and professional growth through teaching, research initiatives, community engagement, and the preservation of cultural values. One key tool in this effort is an educational resource recommendation system, critical in boosting students’ academic outcomes, stimulating engagement in learning activities, and elevating educational standards. This study presents the design and evaluation of a personalized recommendation system for online educational resources, leveraging web technologies and mining techniques integrated within a digital learning infrastructure. The proposed system enhances the intelligence and adaptability of the network-based educational platform, caters to students’ personalized learning requirements, and streamlines the process of accessing relevant information by minimizing the time required for resource discovery.

关键词： Education Decision making Web mining Boosting Research initiatives Global communication Cultural differences Recommender systems Standards

来源：评论

学校读者我要写书评

暂无评论

Patient-Level Anatomy Meets Scanning-Level Physics: Personalized Federated Low-Dose CT Denoising Empowered by Large Language Model

arXiv

引用

arXiv 2025年

作者： Yang, Ziyuan Chen, Yingyu Wang, Zhiwen Shan, Hongming Chen, Yang Zhang, Yi College of Computer Science Sichuan University China Key Laboratory of Data Protection and Intelligent Management Ministry of Education Sichuan University China School of Cyber Science and Engineering Sichuan University China School of Computer Science and Software Engineering Southwest Petroleum University China Institute of Science and Technology for Brain-inspired Intelligence Fudan University China School of Computer Science and Engineering Southeast University China

Reducing radiation doses benefits patients, however, the resultant low-dose computed tomography (LDCT) images often suffer from clinically unacceptable noise and artifacts. While deep learning (DL) shows promise in LDCT reconstruction, it requires large-scale data collection from multiple clients, raising privacy concerns. Federated learning (FL) has been introduced to address these privacy concerns;however, current methods are typically tailored to specific scanning protocols, which limits their generalizability and makes them less effective for unseen protocols. To address these issues, we propose SCAN-PhysFed, a novel SCanning- and ANatomy-level personalized Physics-Driven Federated learning paradigm for LDCT reconstruction. Since the noise distribution in LDCT data is closely tied to scanning protocols and anatomical structures being scanned, we design a dual-level physics-informed way to address these challenges. Specifically, we incorporate physical and anatomical prompts into our physics-informed hypernetworks to capture scanning- and anatomy-specific information, enabling dual-level physics-driven personalization of imaging features. These prompts are derived from the scanning protocol and the radiology report generated by a medical large language model (MLLM), respectively. Subsequently, client-specific decoders project these dual-level personalized imaging features back into the image domain. Besides, to tackle the challenge of unseen data, we introduce a novel protocol vector-quantization strategy Copyright © 2025, The Authors. All rights reserved.

关键词： computerized tomography

来源：评论

学校读者我要写书评

暂无评论

Path Pooling: Train-Free Structure Enhancement for Efficient Knowledge Graph Retrieval-Augmented Generation

arXiv

引用

arXiv 2025年

作者： Wang, Hairu Feng, Yuan Xie, Xike Zhou, S. Kevin School of Computer Science University of Science and Technology of China China School of Biomedical Engineering USTC China Data Darkness Lab MIRACLE Center Suzhou Institute for Advanced Reasearch USTC China

Although Large Language Models achieve strong success in many tasks, they still suffer from hallucinations and knowledge deficiencies in real-world applications. Many knowledge graph-based retrieval-augmented generation (KG-RAG) methods enhance the quality and credibility of LLMs by leveraging structure and semantic information in KGs as external knowledge bases. However, these methods struggle to effectively incorporate structure information, either incurring high computational costs or underutilizing available knowledge. Inspired by smoothing operations in graph representation learning, we propose path pooling, a simple, train-free strategy that introduces structure information through a novel path-centric pooling operation. It seamlessly integrates into existing KG-RAG methods in a plug- and-play manner, enabling richer structure information utilization. Extensive experiments demonstrate that incorporating the path pooling into the state-of-the-art KG-RAG method consistently improves performance across various settings while introducing negligible additional cost. Code is coming soon at https://***/hrwang00/path-pooling. Copyright © 2025, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Internal Audit Effectiveness for data Compliance Using Privacy Federated Learning Model in Hospital Environment

引用

Procedia computer Science 2025年 254卷 171-180页

作者： Chamundeswari Arumugam Kamalika Karuppuchamy C Amarnath B K Jeyasuriya B Department of Computer Science & Engineering Saveetha Engineering College Tamil Nadu India School of Law SASTRA University Thanjavur Tamil Nadu India Data protection consultant Tsaaro consulting Bangalore Karnataka India Engineering GRC Cycatz Technologies Private Limited Bangalore Karnataka India

Corporate governance privacy policies that follow the digital sovereignty, have an impact on firm’s value and it will reflect on M & A. The best practices and access to new technologies of data privacy policies after acquisition will be applied as per GDPR recommendations. The objective of this paper is to apply a privacy model to protect patients as per privacy policies and maintain cyber security data compliance. The checklist of the acquirer cyber security data compliance policy is applied to the acquired hospital by considering the GDPR articles. Federated learning privacy model is applied to protect and preserve the data during training and testing on OCT image dataset. The dataset was split for training and testing, and gaussian noise was added for privacy preservation from inversion attack. Further, to mitigate the label flipping attack, non-linear aggression method is applied and prediction was measured. Internal audit effectiveness for data compliance and noncompliance was explored and multiple linear regression model was applied to check the variance score. The variance score of 0.59 was obtained using the stimulated dataset and further it can be improved by considering the real time dataset.

关键词： computer science Privacy data Compliance Hospital OCT Images Federated Learning Internal audit GDPR digital sovereignty

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：